Schema Documentation¶
This section contains the complete documentation for the RevAIse Data Model schema, automatically generated from the LinkML source definitions.
Schema Structure¶
The RevAIse schema is organized into three main categories:
Main Schema¶
The main schema defines the root Review class and imports all components. This is the entry point for understanding the complete data model.
Review Core Objects¶
These are the fundamental objects that characterize a systematic review:
- Review - The main review container and metadata
- Author - Information about review authors and contributors
- Protocol - Review protocol and registration details
- Literature Record - Individual literature items and their metadata
Shared Infrastructure Objects¶
These objects are imported in review_core.yaml for sharing across stages:
- Registration Template - Templates for review registration
- Stage Execution - Base class for all review stages
- Stage Output - Outputs from stage executions
- Software Environment - Computational environment specifications
- External Tool - External tools and software used in the review
- Enumerations - Controlled vocabularies and value sets
Additional Objects¶
These objects provide additional functionality used by stages:
- Stage Protocol - Base protocol class for all stages
- Stage Quality Control - Base quality control for all stages
- Stage Statistics - Base statistics for all stages
- AI Assistance - AI model configurations and assistance documentation
- Conflict Resolution - Conflict resolution tracking and methods
- Field Definition - Data extraction field definitions and forms
- Participant - Review participants and their roles
- Quality Assessment - Quality assessment tools and metrics
- Quality Metrics - Performance and agreement metrics
- Work Session - Work session tracking and metrics
Review Stages¶
These represent the sequential phases of a systematic review:
- Registration - Protocol registration and pre-registration
- Search - Literature search strategy and execution
- Screening - Title/abstract and full-text screening
- Extraction - Data extraction from included studies
- Synthesis - Data synthesis and meta-analysis
Schema Features¶
Modular Design¶
The schema uses a modular architecture where: - Each component is defined in its own file - Components can be reused across different stages - Extensions can be added without modifying core components
LinkML Benefits¶
Built with LinkML (Linked Data Modeling Language), the schema provides: - Multiple serialization formats - YAML, JSON, RDF, and more - Built-in validation - Automatic generation of validation schemas - Semantic web compatibility - JSON-LD contexts and RDF support - Documentation generation - Auto-generated human-readable documentation - Type safety - Strong typing with ranges and constraints
AI Documentation Support¶
Special attention to documenting AI assistance: - Model specifications and versions - Prompts and parameters used - Human oversight and modifications - Performance metrics and confidence scores
Provenance and Reproducibility¶
Comprehensive tracking of: - Temporal information (dates and durations) - Actor attribution (human and AI agents) - Software environments and tool versions - Decision rationale and modifications
Using the Schema¶
For Developers¶
- Use the JSON Schema for validation in your applications
- Refer to the JSON-LD Context for linked data applications
- Import the LinkML YAML directly for schema extensions
For Researchers¶
- Review the object definitions to understand data requirements
- Follow the stage documentation for process guidance
- Use the enumerations for controlled vocabularies
For Data Managers¶
- Validate data against the JSON Schema
- Ensure all required fields are populated
- Document AI usage according to the schema specifications
Schema Versioning¶
This documentation corresponds to the schema version you're currently viewing. The schema follows semantic versioning:
- Major versions - Breaking changes to the schema
- Minor versions - New features, backward compatible
- Patch versions - Bug fixes and documentation updates
Use the version selector at the bottom of the page to access documentation for different versions.
Quick Navigation¶
| Component | Description | Primary Use |
|---|---|---|
| Main Schema | Complete schema definition | Understanding the full model |
| Review Object | Root review container | Starting a review document |
| Registration Stage | Registration/pre-registration | Protocol documentation |
| Search Stage | Search strategy and execution | Search documentation |
| Screening Stage | Study selection process | Screening documentation |
| Extraction Stage | Data extraction process | Extraction documentation |
| Synthesis Stage | Data synthesis and meta-analysis | Synthesis documentation |
Schema Extensions¶
The RevAIse schema is designed to be extensible. You can: - Add custom fields to existing classes - Create new stage types for domain-specific needs - Define additional enumerations for controlled vocabularies - Extend AI documentation for new model types
For guidance on extending the schema, see the GitHub repository.